Structural Optimization of Turntable Frame Based on Improved Latin Hypercube Design and Hybrid Surrogate Model

Structural optimization is an important part for engineering design. It involves optimizing the product or engineering structure by combining finite element analysis (FEA) with optimization algorithms. Based on the improved Latin hypercube design (LHD) and hybrid surrogate model, a structural optimization method is proposed to optimize the structure. It employs the successive local enumeration (SLE) method to improve LHD and obtain more comprehensive size parameter and sample space information. Based on the FEA results of the sample points, a hybrid surrogate model is established, and it involves polynomial response surface model (RSM), Kriging model, and radial basis function (RBF) neural network model. Thus, the simulation computation time is reduced and more accurate results are obtained. Moreover, the non-dominated sorting genetic algorithm (NSGA-II) is used for multi-objective optimization of the size parameters, and the optimal parameter solution is gained based on the minimum Euclidean distance criterion. Case study of a two-axis turntable frame is conducted to verify the effectiveness of the proposed method. The results show that after the optimization, the first order natural frequency of the turntable frame is reduced by 3.37%, meanwhile the mass is reduced by 8.87% and the maximum deformation is reduced by 19.42%. The results indicate that with minimal loss in the first order natural frequency, significant improvements can be achieved in the properties of mass and deformation.